Most cell lines used for recombinant protein production are chosen based on their ability to proliferate indefinitely. Robust cell growth is essential to clone good producers and to propagate such clones to a high cell density in the production process. Additionally, many previous studies have reported a positive correlation between growth rate and product formation (Smiley et al., 1989, Biotechnol. Bioeng. 33:1182-1190; Cockett et al., 1990, Bio/Technology 8:662-667; Hayter et al., 1991, Appl. Microbiol. Biotechnol. 34:559-564; Robinson and Memmert, 1991, Biotechnol. Bioeng. 38:972-976; Pendse et al., 1992, Biotechnol. Bioeng. 40:119-129), while only few investigations have reported the opposite (Mitchell et al., 1991, Cytotechnology 5:223-231; Bebbington et al., 1992, Bio/Technology 10:169-175; Tonouchi et al., 1992, J. Biotechnol. 22:283-290).
The phenomenon that reduced specific growth rate may increase cell culture productivity of the cell has been observed with hybridoma cells producing monoclonal antibodies (Mab's) (Suzuki and Ollis, 1990, Biotechnol. Prog. 6:231-236; Al-Rubeai et al., 1992, Cytotechnology 9:85-97). However, since in these and other experiments, the G1-arrest of the cell cycle is achieved by starvation of the cells for an essential nutrient or energy source or by the addition of DNA synthesis inhibitors such as thymidine or hydroxyurea, TGF-.beta. (Suzuki and Ollis, 1990, supra; Al-Rubeai et al., 1992, supra) or genotoxic agents such as adriamycin (Gartenhaus et al., 1996, Proc. Nate. Acad. Sci. 93:265-268), such processes are not preferred for use in prolonged production phases since these approaches interfere with cell viability and/or disturb metabolic processes necessary for protein synthesis (Marcus et al., 1985, Ann. Rev. Genet. 19:389-421; Windle et al., 1991, Genes Dev. 5:160-174; Di Leonardo et al., 1993, Cold Spring Harbor Symp. Quant. Biol. 58:655-667; Wertz and Hanley, 1996, TIBS 21:359-364, for a review).
The first relevant biotechnological contribution to protein production in proliferation-inhibited cells came from the isolation of a temperature-sensitive CHO cell line. These mutants showed 3 to 4 times higher production of a tissue inhibitor of metalloproteinases (TIMP) when growth arrest was induced by a temperature shift to 39.degree. C. (Jenkins and Hovey, 1993, Biotechnol. and Bioeng. 42:1029-1036). Although the productive life span of the cells was extended in arrested cultures and the overall productivity was higher, the mutant cells showed a rapid decrease in viability upon prolonged exposure to elevated temperatures. Thus, while temperature-dependent growth control can be an attractive approach from a process viewpoint, potential problems include not only viability loss at elevated temperatures but also potential lower productivity at suboptimal temperatures.
Most successful cell culture processes make use of proliferating cells. Inevitably, proliferation beyond a certain desired cell density causes nutrient and oxygen depletion, accumulation of lactate and other toxic products, and deterioration and degradation of the product. Much efforts are devoted to improve positive control of cultured cell proliferation, e.g., by replacing the growth factor-containing animal serum by more sophisticated, better-defined technologies which involve use of defined chemical media with defined protein additives and/or adaptation of cell lines, or their genetic engineering, to permit growth in low- or no-protein medium (Renner et al., 1995, Biotech. Bioeng. 47:476-482; Lee et al., 1996, Biotech. Bioeng. 50:273-279). For example, engineering cells to overexpress survival genes from the bcl-2 family, such as bcl-2, bcl-x.sub.L and E1B, has been used for higher robustness and obviation of the cellular death program in production. (Boise et al., 1993, Cell 74:597-608; Chiou et al., 1994, Molecular and Cellular Biology 14:2556-2563; Itoh et al., 1995, Biotechnol. Bioeng. 48:118-122; Cherbonnel-Lasserre et al., 1996, Oncogene 13:1489-1497; Han et al., 1996, Genes and Dev. 10:461-477; Murray et al., 1996, Biotech. Bioeng. 51:298-304; Singh et al., 1996, Biotechnol. Bioeng. 33:1182-1190; Huang et al., 1997, Oncogene 14:405-414).
Production using growing cells contrasts with the natural situation in which many cells of a mature mammal display no further growth after terminal differentiation yet continue to produce and often secrete proteins during the lifetime of the organism. In an intact mammalian organism, uncontrolled proliferation is detrimental and can result in tumor production.
Many proteins which inhibit proliferation have emerged from cancer research where the absence or mutation of the respective genes corresponds to a state of uncontrolled proliferation and cancerous growth. Such genes are referred to as tumor suppressor genes. The p53 tumor suppressor gene is the most commonly mutated gene in human cancer, with the majority of mutations being amino acid substitutions (Lane, 1992, Nature 358:15-16; Greenblatt et al., 1994, Cancer Res. 54:4855-4878; Ko and Prives, 1996, Genes Dev. 10:1054-1072).
The normal role of p53 is to induce cell-cycle arrest predominantly at the G1-checkpoint in response to DNA damage by binding to damaged DNA in a non-specific fashion via its C-terminal domain (Kastan et al., 1991, Cancer Res. 51:6304-6311). p53 has also been implicated as a general metabolic sensor (Linke et al., 1996, Genes Dev. 10:934-947). The G1-checkpoint control function is executed by accumulation of p53 followed by the site-specific binding to promoter elements and corresponding induction of the MDM2, GADD45, IGF-BP3 and p21 genes (Kastan et al., 1991, supra; Barak et al., 1993, EMBO J. 12:461-468; El-Deiry et al., 1994, Cancer Res: 54:1169-1174).
Induction of p53 leads not only to cell growth arrest but also to programmed cell death, or apoptosis, under some conditions. Both the cytostatic and apoptosis mechanisms enable p53 to control DNA damage by protecting cellular descendants from accumulating excessive mutations (see Ko and Prives, 1996, supra, for a review). The molecular process determining whether cells subjected to genotoxic stress arrest their growth for subsequent repair of DNA damage or undergo apoptosis is poorly understood. The overall response to p53 expression varies among different cell lines (Yonish-Rouach et al., 1991, Nature 352:345-347; Shaw et al., 1992, Proc. Natl. Acad. Sci. 89:4495-4499; Liu et al., 1994 and 1995, Cancer Res. 54:3662-3667, 55:3117-3122; Symonds, et al., 1994, Cell 78:703-711; Yang et al., 1995a, Cancer Res. 55:4210-4213; Polyak et al., 1996, Genes Dev. 10:1945-1952). Recently, a p53 mutant designated p53175P has been isolated from a cervical carcinoma (Crook et al., 1994, Cell 79:817-827); this mutant showed a specific loss of apoptotic but not cell-cycle arrest function (Rowan et al., 1996, EMBO J. 15:827-838).
p21, another tumor suppressor gene, inhibits DNA replication and enhances DNA repair by interaction with the replication and repair factor, the proliferating cell nuclear antigen designated PCNA (Xiong et al., 1993a and b, Nature 366:701-704, Genes Dev. 7:1572-1583; Flores-Rozas, et al., 1994, Proc. Natl. Acad. Sci. 91:8655-8659; Smith et al., 1994, Science 266:1376-1380; Waga et al., 1994, Nature 369:574-578). At high concentrations, p21 also inhibits the function of cyclin-dependent kinases (Cdks), particularly those that function during the G1-phase of the cell cycle (Gu et al., 1993a, Biotechnol. Bioeng. 42:1113-1123; Harper et al., 1993, Cell 75:805-816; Xiong et al., 1993a, supra). These Cdks normally phosphorylate the product of the retinoblastoma tumor suppressor gene (Rb) in a cyclin-dependent manner which subsequently activates the E2F family of transcription activators and eventually leads to cell-cycle progression and the beginning of S-phase (Pagano et al., 1992, Science 255:1144-1147; Lees et al., 1992, Genes Dev. 6:1874-1885).
Besides activation by p53, p21 can also be induced by the CCAAT/enhancer-binding protein .alpha. (C/EBP.alpha.) as part of the cell-cycle inhibition program following terminal differentiation (Timchenko et al., 1996, Genes & Dev. 10:804-815). Additionally, C/EBP.alpha. stabilizes p21 at the protein level and increases thereby its half-life (Timchenko et al., 1996, supra). Recently, it has also been reported that mutant fibroblast cells lacking a functional interferon responsive factor gene (IRF-1) do not express p21 following cytotoxic stress (Tanaka et al., 1996, Nature 382:816-818). IRF-1 is a DNA-binding transcription activator which accumulates in response to interferons and leads to the general antiviral state of the cells (Pine 1992, J. Virol. 66:4470-4478). Regulated overexpression of IRF-1 was shown to reversibly control proliferation in mammalian cells (Koster et al., 1995, in "Animal Cell Technology: Developments Towards the 21st Century", pp. 33-43, E. C. Beuvery et al. (eds.), Kluwer Academic Publ., Dordrecht, The Netherlands), but, unlike p21, IRF1-mediated growth arrest does not lead to the inhibition of proliferation at a specific cell cycle phase (Kirchhoff et al., 1995, Oncogene 11:439-445). Additionally, induction of IRF-1 leads to a decrease in cell viability and a decrease in production of a marker gene unless that marker gene is placed under the direct transcriptional regulation of the IRF-1 DNA binding factor (Kbster et al., supra).
Another pathway, which is p53-independent, blocks the mammalian cell cycle in G1 in response to the anti-mitogenic cytokine called transforming growth factor-.beta. (TGF-.beta.). p27, a cyclin-dependent kinase inhibitor (Cdi) is responsible for this cell-cycle block (Coats et al., 1996, Science 272:877-880). p27 and p21 are highly homologous in their N-terminus, suggesting that they may bind and inhibit cyclin-Cdk complexes in a similar fashion (Polyak et al., 1994a, Cell 78:59-66; Toyoshima and Hunter, 1994, Cell 78:67-74). However, p21 and p27 differ in their affinity for different Cdks; p21 binds most tightly to Cdk2, whereas p27 binds with more affinity to Cdk4 (Polyak et al., 1994a, supra). Moreover, unlike p21, which is primarily regulated at the level of transcription, the level of p27 mRNA and protein remains the same in quiescent and proliferating cells, and throughout the cell cycle (Toyoshima and Hunter, 1994, supra). p27 is regulated through binding a heat-labile "masking" factor, and is unmasked when cells are treated with TGF-.beta., or become contact-inhibited (Polyak et al., 1994b, Gene and Dev. 8:9-22; Sherr, 1994, Cell 79:551-555). Recently, p21 has also been shown to be induced by TGF-.beta. and to cooperate with Ink4 Cdk inhibitors to induce p53-independent cell-cycle arrest (Datto et al., 1995, Proc. Natl. Acad. Sci. 92:5545-5549; Reynisdottir et al., 1995, Genes and Dev. 9:1831-1845). However, p53 can also indirectly interfere with Cdk4 inhibition by repressing translation of the respective mRNA (Ewen et al., 1995, Genes Dev. 9:204-217).